I recently bought a G10/241E tube and I need a power supply for it. There are no designs of a 330V DC power supply and no buck-boost converters would work with such voltage. The datasheet and example circuit:
Unidirectional Cold-Cathode Gas-Filled Decade Counter - G10/241E (CVX2223)
Is there any easy boost converter or should I just buy a transformer and power it from mains? (I have 220V mains.)
It's common to run these things from the mains supply. It can be done like so:
simulate this circuit – Schematic created using CircuitLab
I've skipped some of the detail, but essentially two back-to-back transformers will do it. Pay attention to the way they are arranged. You'll want some smoothing and filtering things after the bridge, and possibly some fusing on the mains side. The transformers can be little things (cheap) as you're not using any significant current.
If you allow 75% losses (anecdotal), you'll get about 345V DC. That's within spec from your data sheet. I can't be more specific as the losses can't be calculated more precisely. You'll have to spend a little to try it. You might be able to make adjustments by varying the step down ratios of the transformers. Another way is to just drop the voltage through a large power resistor. You wouldn't be dropping much, so its size should be reasonable and this is done all the time in valve amp PSUs. Some of the filtering can be accomplished with this resistor anyway.
I found a reference to ball parking the losses of a back to back transformer arrangement. You'll see that theoretically you'd expect 120V back out again, but the guy only got 93V. So expect an efficiency of 78%. Perhaps some voltage disappeared due to heater loading, but the only way'll you find out for sure in your particular circumstances is to experiment.
.
I did something similar in my lab at home .12V in and 220VDC out for testing Valve car radios used in Classic cars .I used a simple boost converter with a switching loss reduction scheme.This allowed a much larger boost ratio to be supported with approx 90% efficiency . The lash up was a through hole birds nest .When I used similar circuits commercially I specified a Wurth SMD boost coil because it had a 400 Volt peak rating .These days it is much cheaper and easier to use a SMD coil than to wind a transformer so it was very worthwhile to get these big boost ratios right.I used a 400 V TO220 mosfet for the car radio birds nest .The peak voltage rating of the available coil is the limiting factor here but you will be fine because you only want 330V.For voltages in the KV range at low currents I have used diode pumps of various numbers of stages keeping peak drain voltages low .The S Trap boost converter runs well into a diode pump and can happily make positive and negative voltages .You could get your favorite boost converter and just try it on a diode pump keeping the boost ratio low so your switching losses are not too bad .Most other SMD coil manufacturers do not state peak voltage rating but if you are only boosting to say 60V and pumping the rest then any cheap SMD coil will do .My circuit uses discretes because no chips at present give me the desired switching regime for the switching loss reduction scheme.What I am saying is that for you a chip could be valid because the voltage requirements are lower and at such low powers component footprints dictate size rather than heat dissipation.
